You read a system's solution off its graph by testing points and reading boundary lines. Find where the shaded regions overlap, then confirm a sample point satisfies every inequality at once. Solid boundaries include their line; dashed boundaries exclude it. That overlap region plus the matching boundary segments is the solution set. This holds for two-variable systems at Algebra I level and for linear programming models, though a graph alone leaves coordinates approximate unless a vertex lands on a grid intersection.
Read Solid and Dashed Boundaries Before Anything Else
Every inequality in a two-variable system gets graphed as a boundary first. For a strict inequality (y > 2x + 1), you draw a dashed line because points on that line make both sides equal, not greater. For a non-strict inequality (y β₯ 2x + 1), you draw a solid line because the boundary itself satisfies the relation. The line is never the whole answer β it's a fence, and you still have to decide which side to stand on.
Take y < βx + 4 with a dashed line through (0, 4) and (4, 0). The point (0, 0) gives 0 < 4, which is true, so shade the side containing the origin. Substitute the origin into every inequality you meet, because (0, 0) makes arithmetic trivial. When the boundary passes through the origin, pick (1, 1) or (0, 1) instead. Write your test-point result next to the line. That note saves you from re-testing when you revisit the graph.
The single most common error is flipping the shading after multiplying by a negative. If you rearrange β2y > 6 to y < β3, the inequality direction reverses; shade below y = β3, dashed. Students who shade above get a region that satisfies neither the original inequality nor the rewritten one. Test the origin after every sign flip.
Dashed lines are excluded not because the algebra forbids them but because equality isn't part of the relation. That distinction matters at the overlap: if one boundary is dashed and the other solid, the shared edge is only partly included β the segment on the dashed line is excluded while the segment on the solid line is included. Mark included segments with a thick pen and excluded ones with an open circle at each vertex, then you won't misread the final region.
Isolate y So You Can Shade Without Recomputing
Putting each inequality in slope-intercept form (y > mx + b or y β₯ mx + b) makes shading mechanical. Once y is isolated, "greater than" means above the line and "less than" means below, with no test-point arithmetic. For 3x β 4y β€ 12, subtracting 3x gives β4y β€ β3x + 12, and dividing by β4 flips the sign to y β₯ (3/4)x β 3. Shade above a solid line through (0, β3) and (4, 0).
The flip rule applies only to multiplication or division by a negative number. Adding or subtracting a term across the inequality never changes direction. If you're ever unsure, test a point after rearranging: y β₯ 0.75x β 3 must hold at (0, 0) because 0 β₯ β3, and the origin sits above the line, confirming "above" is correct.
Systems with x isolated or with y on both sides need the same treatment. For 2x + 3y < 6x β y + 9, collect terms: 4y < 4x + 9, so y < x + 2.25, dashed, shade below. Note the decimal in the intercept β many textbooks keep fractions (9/4) for exactness, and you should too when a vertex answer depends on the intersection.
Slope-intercept form also reveals parallel boundaries instantly. If both inequalities simplify to the same slope but different intercepts, the lines never cross and the overlap is either a slab (between them) or empty. Recognising that before you shade saves a full graph's worth of work and prevents the misleading "shade everything either side" mistake that turns a no-solution system into a false region.
Locate the Overlap Region Using Three Test Points
The solution to a system is the intersection of the individual solution sets β every point that satisfies all inequalities simultaneously. On a graph with two shaded half-planes, the answer is the doubly shaded area. Testing three points confirms which area that is: one inside the overlap, one inside only a single region, and one outside both. If the inside point satisfies all inequalities and the other two fail at least one, the overlap is identified.
Work a concrete example. System: y β₯ x + 1 (solid, above) and y β€ βx + 5 (solid, below). They intersect at (2, 3). Test (2, 3): satisfies both with equality. Test (0, 2): 2 β₯ 1 true, 2 β€ 5 true, so inside the overlap. Test (4, 2): 2 β₯ 5 false, so outside, even though it satisfies the second inequality. The overlap is the wedge between the lines, extending upward and downward until the boundaries cross.
| Test point | y β₯ x + 1 | y β€ βx + 5 | Verdict |
|---|---|---|---|
| (2, 3) | 3 = 3 β (boundary) | 3 = 3 β (boundary) | Vertex, included |
| (0, 2) | 2 β₯ 1 β | 2 β€ 5 β | Inside overlap |
| (4, 2) | 2 β₯ 5 β | 2 β€ 5 β | Outside |
| (3, 7) | 7 β₯ 4 β | 7 β€ 2 β | Outside |
Read the table in one pass if you like β the β/β columns tell you which region to shade without drawing to scale. Three test points suffice for two-inequality systems; three-inequality systems need a test point per region created by the lines, which is at most seven for three lines in general position. When you find an inside point, shade the region containing it, not the region containing the origin. The origin is convenient, but not sacred.
Shade Correctly for β₯, β€, >, and < Rules
Shading follows the direction the inequality points after isolation. y > β¦ means above; y < β¦ means below. But mixed systems sometimes isolate x instead, and then "greater than" means to the right and "less than" to the left, which many students miss. For x > 2 (vertical dashed line), shade right of the line; x β€ β1 shades left of the solid line.
Horizontal lines behave analogously. y β₯ 3 shades above the solid line at height 3; y < β2 shades below the dashed line at height β2. Combining a vertical and horizontal constraint produces a quadrant: x β₯ 0, y β₯ 0 is the first quadrant including the axes, and the solution to that system alone is an unbounded quarter-plane. Add x + y β€ 10 and you've defined the standard linear-programming feasible region: a triangle with vertices (0, 0), (10, 0), and (0, 10).
Boundary inclusion changes how you report the answer. A region shaded above a dashed line means the answer is written with a strict sign (y > β¦) on the boundary segment. If the same region sits under a solid line, the answer carries β₯ on that segment. When you write the final solution as a description like "the region satisfying 1 β€ x + y β€ 5, x β₯ 0, y β₯ 0", you've replaced the graph with algebra and settled whether the boundary counts.
Gradient or colored shading is your friend with three or more inequalities. Shade each region once with a distinct color at about 30 percent opacity; the region covered by every color is the answer. Four inequalities yield four overlapping color layers, and points inside all of them satisfy all of them.
Treat Intersections as Candidate Vertices, Not the Whole Answer
The boundary lines' intersection points are the vertices of the solution polygon. In the system y β₯ x + 1 and y β€ βx + 5, the intersection (2, 3) is a vertex, but the solution set itself extends infinitely upward β it isn't just a point. The vertex is where the two boundaries meet; the solution is everything inside the wedge between them.
Three or more inequalities produce a bounded polygon, and you find its vertices by solving each pair of boundary equations. Take y β₯ 0, x β₯ 0, and x + y β€ 6. The boundary pairs intersect at (0, 0), (6, 0), and (0, 6), giving a right triangle. Each vertex comes from a 2Γ2 system: x = 0 with x + y = 6 gives (0, 6); y = 0 with x + y = 6 gives (6, 0); x = 0 with y = 0 gives (0, 0). Solve systematically, not by eye.
The graph shows whether a candidate vertex belongs to the solution. Test (6, 0) in all three inequalities: 0 β₯ 0 β, 6 β₯ 0 β, 6 β€ 6 β, so the vertex is included because every boundary is solid. If x + y β€ 6 were x + y < 6 instead, the same triangle would exclude the hypotenuse and the vertices (6, 0) and (0, 6) would be open. The vertex positions don't change; inclusion does. That is why dashed lines deserve special attention at corners.
Vertices matter beyond description because they're where objective functions are optimised. A linear program such as maximise P = 3x + 2y subject to a triangle has its maximum at a vertex by the corner-point principle. Evaluate P at (0, 0) = 0, (6, 0) = 18, and (0, 6) = 12; the max is 18 at (6, 0). A graph alone gives approximate vertices, so verify the 2Γ2 system algebraically before reporting an optimum.
Four-Inequality Systems: Walk the Feasible Region in Order
A standard linear programming setup such as x + y β€ 8, x + 2y β€ 12, x β₯ 0, y β₯ 0 produces a quadrilateral feasible region. Find the vertices by solving consecutive constraint pairs. The boundary lines x + y = 8, x + 2y = 12, x = 0, y = 0 intersect at (0, 0), (8, 0), (4, 4), and (0, 6). Checking each vertex in all four inequalities confirms inclusion, and (6, 1) is outside because 6 + 2(1) = 8 β€ 12 β but 6 + 1 = 7 β€ 8 β actually β check carefully: (6, 1) is inside both sums, so it's inside the region, not outside. The vertex (8, 0) has x + 2y = 8 β€ 12 β and x + y = 8 β€ 8 β, included. The vertex (0, 6) has x + 2y = 12 β€ 12 β, included.
Work through the region edge by edge to avoid the slip above. From (0, 0) along the x-axis to (8, 0), then along x + y = 8 until it meets x + 2y = 12 at (4, 4), then along x + 2y = 12 to the y-axis at (0, 6), and back down the y-axis to (0, 0). Each edge is one constraint held at equality. The edge from (0, 6) to the origin is x = 0. The edge from (4, 4) to (0, 6) is x + 2y = 12. There are exactly four edges, one per constraint, because each is active on one side of the polygon.
| Vertex | x + y β€ 8 | x + 2y β€ 12 | x β₯ 0 | y β₯ 0 | Status |
|---|---|---|---|---|---|
| (0, 0) | 0 β€ 8 β | 0 β€ 12 β | 0 β₯ 0 β | 0 β₯ 0 β | Included |
| (8, 0) | 8 β€ 8 β | 8 β€ 12 β | 8 β₯ 0 β | 0 β₯ 0 β | Included |
| (4, 4) | 4 β€ 8 β | 4 β€ 12 β | 4 β₯ 0 β | 4 β₯ 0 β | Included |
| (0, 6) | 6 β€ 8 β | 6 β€ 12 β | 0 β₯ 0 β | 6 β₯ 0 β | Included |
| (6, 1) | 6 β€ 8 β | 8 β€ 12 β | 6 β₯ 0 β | 1 β₯ 0 β | Included (interior) |
The table is your audit. When every vertex of a candidate polygon passes all constraints, the polygon is the feasible region. When a candidate vertex fails, the polygon is the wrong shape or the system has no overlap. Interior points like (6, 1) test constraints but don't define vertices; report vertices for objective-function work and interior points only as confirmation.
Add a fifth constraint, say 2x + y β€ 10, and see whether it cuts the quadrilateral. At (4, 4) the new expression equals 12 > 10, so (4, 4) becomes infeasible and the region shrinks to a pentagon with a new vertex where 2x + y = 10 meets x + 2y = 12. Dividing 2x + y = 10 by 2 gives x + 0.5y = 5; subtracting from x + 2y = 12 gives 1.5y = 7, so y = 14/3 and x = 8/3. The pentagon vertices are (0, 0), (5, 0), (8/3, 14/3), (4, 4)β¦ except (4, 4) is now outside. Recompute: (4, 4) gives 12 > 10, infeasible. The pentagon is (0, 0), (5, 0), (8/3, 14/3), (0, 6). A graph alone would hide the fractional vertex unless you solve the pair algebraically.
Detect No-Solution and Infinite-Solution Graphs
Parallel boundaries with opposite shading create no solution. y > 2x + 1 and y < 2x β 3 both have slope 2. The first shades above y = 2x + 1, the second below y = 2x β 3. Since 2x β 3 sits below 2x + 1 for every x, the two regions face away from each other and share nothing. The graph shows two parallel dashed lines with shading pointing outward, and the answer is the empty set, written β or "no solution". Testing any point confirms it: (0, 0) gives 0 > 1 false and 0 < β3 false, so the origin belongs to neither region, and no point can belong to both.
Infinite solutions arise when two inequalities describe the same half-plane or when one contains the other. y β₯ x and y β₯ x + 1: the second is stronger, so the first is redundant and the solution is everything on or above y = x + 1. The graph shows both lines, but the answer is a single unbounded region. y > x and y β₯ x describe the same region β the union of "above" with the line yields β₯ β so the system has infinitely many solutions in one half-plane.
A subtler case is a system with infinitely many solutions along a line. y = x + 1 (an equation, not an inequality) together with y β₯ x + 1 gives the single ray y = x + 1 for x β₯ 0 if you also impose x β₯ 0. The overlap is a ray, not a region. Report the ray, not the whole line.
Every empty solution has a witness point that violates at least one inequality, and you can name it. Every infinite solution has a line or a pair of parallel lines explaining the redundancy. Test the deletion: remove one inequality; if the answer doesn't change, it was redundant. If the answer expands to all real numbers, the remaining inequalities are inconsistent only with each other.
Check Your Answer Against the Original Inequalities
Graphing can slip. A shading error, a misread intercept, or a sign flip can produce a region that looks right but fails algebra. Verify by taking three points inside your final region and substituting them into the original, un-simplified inequalities. Any failure means the graph is wrong. For y β₯ 0.75x β 3, test (0, 0): 0 β₯ β3 β. Test (4, 0): 0 β₯ 0 β, so the x-axis point (4, 0) is on the boundary. Test (8, 2): 2 β₯ 3 false, so (8, 2) is outside β good, because it should be.
Verification also catches vertex-position errors. If you found (4, 4) as a vertex but recomputing the 2Γ2 system gives (4, 4) β you already had it β the check is trivial. But if the graph shows a vertex near (4.1, 3.9), the algebra wins. Re-solve the pair: x + y = 8 and x + 2y = 12 give y = 4, x = 4. Any point labelled (4.1, 3.9) on a hand-drawn graph is an artefact of the drawing, not the mathematics.
Run a sign audit before reporting the answer. Recheck every multiplication or division by a negative, then recheck the shading direction it produced. Two-thirds of graphing mistakes trace to one flipped sign. Confirm strict inequalities use dashed boundaries and non-strict use solid. Confirm axes-signed constraints (x β₯ 0, y β₯ 0) are included unless the context (a length in a word problem, for example) forces different boundaries.
Turn Graphs Into Linear Programming Answers
A graphical solution is a picture, but fields like operations research demand numbers. The corner-point method converts the picture into an answer by evaluating an objective function at each vertex. Maximise 5x + 4y subject to x + y β€ 8, x + 2y β€ 12, x β₯ 0, y β₯ 0. The vertices are (0, 0), (8, 0), (4, 4), (0, 6). Evaluate: 0, 40, 36, 24. The max is 40 at (8, 0).
Verify the answer isn't empty. If constraints also required y β₯ 5, then x + y β€ 8 and y β₯ 5 allow only y between 5 and 8 with x = 0β3, and (8, 0) becomes infeasible. Re-evaluate: with y β₯ 5, the vertices become (0, 5), (0, 6), and (3, 5). Evaluate 5x + 4y: 20, 24, 35. The max is 35 at (3, 5). A graph alone wouldn't warn you the previous best point was infeasible under the new constraint.
Different objectives produce different optima on the same region. Maximise 2x + 3y on the same triangle (0, 0), (6, 0), (0, 6): values 0, 12, 18, so (0, 6) wins. Maximise 3x + 2y: 0, 18, 12, so (6, 0). The optimum moves with the coefficients, which is why the vertices must be exact, not eyeballed. Solve each pair algebraically, keep fractions, and only convert to decimals at the end.
When a vertex is excluded by a dashed line, the maximum may not exist. Maximise x + y over y < x and y > x β 2: the region is an open strip 0 < x β y < 2, and x + y is unbounded above. The graph shows the unshaded boundary lines, and the LP is unbounded. Corner-point methods require the feasible region to be closed and bounded for a guaranteed maximum β check the shape before optimising.
FAQ
What does it mean if my shaded regions don't overlap at all?
The system has no solution, written β or "no solution". Parallel boundaries with opposing directions of shading are the classic cause, but any pair of half-planes can miss each other if their boundaries cross outside the region you've drawn. Test the midpoint of the line segment between any two candidate points on each inequality; if that midpoint violates one inequality, the two shadings have no overlap.
Can a system of inequalities have exactly one solution point?
Only in degenerate cases, not with two-variable strict inequalities. A single solution arises for a system like y β₯ x, y β€ x, y β₯ 0, x β₯ 0 that pins a single point only if all four bind at once, which requires the boundaries to intersect at the origin in a way they usually don't. Practically, one-point solutions come from inconsistent systems with redundant boundaries or from mixed equality systems (y = x + 1 and y β₯ x + 1 and x + y = 3, which intersect at (1, 2)).
How do I know which side of a line to shade without plugging in the origin?
Isolate y so the inequality is y > mx + b or y < mx + b. If y is greater, shade above; if less, shade below. If x is isolated instead, greater means right and less means left. When the inequality is not in isolated form, plug in any convenient point (origin if it's not on the line) and test. The test-point rule never fails, but the form rule saves time on exams.
Does a dashed boundary ever contain points that are part of the answer?
No β dashed boundaries are excluded by definition. If the inequality is strict (for example y > 2x + 1), points on y = 2x + 1 make the inequality false because equality isn't included. However, when two dashed boundaries intersect, the intersection point is also excluded. If a solid boundary crosses a dashed boundary, the vertex on the solid boundary may still be included if it satisfies the other inequality.
What's the difference between solving a system of equations and a system of inequalities from a graph?
An equation system's solution is a set of points (or a line if dependent); an inequality system's solution is an entire region, with boundary inclusion depending on strictness. Graphically, equations give intersections (points), inequalities give overlaps (areas). The algebra is the same substitution/elimination work, but the answer format shifts from coordinates to region descriptions with inequality signs.
At what grade level does this topic typically appear?
Graphing systems of inequalities is standard in Algebra I in most U.S. curricula, typically in the second semester, though linear programming applications appear in Algebra II and precalculus. Region questions account for a substantial share of the systems unit on many state assessments, and commercial test-prep platforms like Khan Academy and IXL place the topic in their Algebra I scope. Check your local standards (Common Core A-REI.D.12, for example) for exact placement.
Why does my vertex come out as a decimal when the lines cross cleanly on paper?
The decimal is an artefact of a graph that isn't perfectly to scale, not a rounding of the algebra. Solve the pair of boundary equations exactly: for y = 2x + 1 and y = βx + 5, setting the right sides equal gives 2x + 1 = βx + 5, so 3x = 4 and x = 4/3, y = 11/3. The vertex is a fraction, and the drawing approximates it. Keep fractions for the final answer and convert to decimals only if a problem demands them.
Do I need to shade every inequality, or can I shade just the overlap?
Shade every inequality individually first, then identify the overlap. If you only shade the overlap, you can't show your reasoning and you can't detect a shading error. For three or four inequalities, use distinct light colours or hatch patterns at low opacity; the darkest region (covered by all) is the feasible set. A single dark region drawn directly is faster but hides mistakes from you and from anyone grading the work.